[PR] Fix decimal fixed-point multiplication for large scale reductions [arrow-rs]

["skyon0522-ai (via GitHub)" <gi...@apache.org>]

skyon0522-ai opened a new pull request, #11708:
URL: https://github.com/apache/arrow-rs/pull/11708

   # Which issue does this PR close?
   
   - Closes #11605
   
   # Rationale for this change
   
   Fixed-point Decimal128 multiplication can return nonzero values for zero products when reducing the scale by 77. Larger reductions, or sufficiently negative input-scale sums, can also overflow the intermediate scale arithmetic in debug builds.
   
   # What changes are included in this PR?
   
   Compute scales in `i16` and explicitly return rounded zeros when the scale reduction exceeds 76. Every product of two `i128` coefficients has magnitude at most 2^254, less than half of 10^77, so all such products round to zero. The existing checked and unchecked arity helpers preserve nulls, length errors, and output metadata. Smaller reductions keep the existing rounding path.
   
   # Are these changes tested?
   
   New regressions exercise reductions of 77 and 128 through the checked, unchecked, and dynamic entry points, including nullable, non-null, empty, and unequal-length inputs. A boundary test retains the nonzero results at reduction 76. Negative-scale tests cover product-scale sums below -128 and direct multiplication at -128; existing rounding tests remain unchanged.
   
   # Are there any significant user-facing changes?
   
   No public API change.
   
   AI assistance was used for source analysis, the implementation, regression tests, and this PR draft.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

Re: [PR] Fix decimal fixed-point multiplication for large scale reductions [arrow-rs]

["alamb (via GitHub) via github" <gi...@arrow.apache.org>]

alamb commented on PR #11708:
URL: https://github.com/apache/arrow-rs/pull/11708#issuecomment-6533163172

   Thank you for your interest in contributing to arrow-rs.
   
   I am closing this PR because it matches the pattern of submissions covered by our [AI generated submissions policy](https://github.com/apache/arrow-rs/blob/main/CONTRIBUTING.md#ai-generated-submissions): a high volume of PRs across many repositories, opened very shortly after issues are filed, where it is not clear that the author is able to debug and own the changes themselves.
   
   I am sorry if this isn't AI spam, but we are being overwhelmed with such PRs and won't have time to review these. Maintainer review time is the scarcest resource in this project, and we need to spend it on contributions where the author can work through the review with us.
   
   If you would like to contribute in the future, please pick a single issue, make sure you fully understand the change and can respond to review comments, and open one PR at a time.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

Re: [PR] Fix decimal fixed-point multiplication for large scale reductions [arrow-rs]

["alamb (via GitHub) via github" <gi...@arrow.apache.org>]

alamb closed pull request #11708: Fix decimal fixed-point multiplication for large scale reductions
URL: https://github.com/apache/arrow-rs/pull/11708


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org